Google Cloud monitored resources

Edit online

After validating your targets, Turbonomic updates the supply chain with the entities that it discovered. The following table describes the entity mapping between the target and Turbonomic.

Google Cloud Turbonomic
Virtual Machine (VM) instance Virtual Machine (VM)
Managed instance group (MIG) Virtual Machine Spec
Unmanaged instance group Not shown as an entity in the supply chain but added to the Groups category in the Search page
Disk Volume
Zone Zone
Region Region

Points to consider:

  • Google Cloud projects, folders, and billing accounts do not appear as entities in the supply chain. Use Search to scope to these resources. In Search, projects are grouped under Accounts, folders under Folders, and billing accounts under Billing Families.

  • Turbonomic supports discovery and management of entities in certain Google Cloud regions. For details, see Supported Google Cloud regions.

Monitored resources for virtual machines

Note:

For a list of supported instance types for virtual machines, see Supported Google Cloud instance types.

Turbonomic discovers Google Cloud labels attached to VMs as tags. You can filter VMs by tags when you use Search or create groups. The Action Details page for a pending VM action also lists all the discovered tags.

Turbonomic monitors the following resources:

  • Virtual memory (vMem)

    Virtual memory (vMem) is the measurement of memory that is in use.

    Turbonomic can collect memory metrics through the following Google Cloud agents:

    • Ops Agent (recommended)

      Install and configure Ops Agent on each VM that Turbonomic monitors. See the Google Cloud documentation for Ops Agent installation instructions and configuration details.

    • Legacy monitoring agent (not recommended)

      For the latest information about the legacy monitoring agent, see the Google Cloud documentation. Turbonomic aligns with Google Cloud's level of support for the agent.

  • Virtual CPU (vCPU)

    Virtual CPU is the measurement of CPU that is in use.

    Turbonomic calculates CPU based on the normalized CPU frequency and the number of vCPUs for a given VM. Normalized CPU frequency takes into account performance variations seen in different models of a given CPU platform. Because frequency is normalized, charts might show utilization values that are slightly higher than 100% (for example, 100.03%) when capacity is fully utilized.

  • Net throughput

    Net throughput is the rate of message delivery over a port.

  • Storage amount

    Storage amount is the measurement of storage capacity that is in use.

  • Storage access (IOPS)

    Storage access, also known as IOPS, is the per-second measurement of read and write access operations on a storage entity.

  • I/O throughput

    I/O throughput is the measurement of an entity's throughput to the underlying storage.

    For both storage access (IOPS) and I/O throughput, Turbonomic calculates capacity or uses capacity data published by Google Cloud, depending on the VM's machine type and disk.

    • Shared-core machine types share a physical core and are used for running small, non-resource intensive apps.

      For shared-core machine types with standard disks or SSDs, Turbonomic calculates capacity using internal benchmark data, which takes into consideration the observed maximum limit that can be achieved for IOPS and I/O throughput. It then uses the calculated capacity to analyze utilization more accurately.

      The use of internal benchmark data could result in differences in the capacity data shown in Turbonomic and Google Cloud.

    • For machine types that are not shared-core:

      • Turbonomic uses published capacity data and assumes that I/O block size is 16KB per I/O.

      • For machine types with persistent disks, Turbonomic assumes that the published capacity for the SSD disk type also applies to the balanced and extreme disk types. When a VM is attached to at least one of these disk types, capacity is assumed to be the per-VM limit for the SSD disk type. When a VM is attached only to the standard disk type, capacity is the per-VM limit for the standard disk type.

Monitored resources for volumes

For a list of supported volume types, see Supported Google Cloud volume types.

Turbonomic monitors the following resources:

  • Storage amount

    Storage Amount is the storage capacity (disk size) of a volume.

    Turbonomic discovers Storage Amount, but does not monitor utilization.

    For a Kubeturbo (container) deployment that includes volumes, Kubeturbo monitors Storage Amount utilization for the volumes. You can view utilization information in the Capacity and Usage chart.

  • Storage access (IOPS)

    Storage access, also known as IOPS, is the measurement of IOPS capacity that is in use.

  • I/O throughput

    I/O throughput is the measurement of I/O throughput capacity that is in use.

Note:

Turbonomic also monitors the attachment state of volumes and generates delete actions for unattached volumes.

Monitored resources for virtual machine specs (MIGs) and unmanaged instance groups

Turbonomic monitors the following resources:

  • Virtual memory (vMem)

    Virtual memory (vMem) is the measurement of memory that is in use.

    For a managed or unmanaged instance group, vMem represents the memory used by all VMs in the group.

    Turbonomic can collect memory metrics through the following Google Cloud agents:

    • Ops Agent (recommended)

      Install and configure Ops Agent on the instance template for the MIG. This strategy automatically installs Ops Agent as new VMs are created in the MIG.

      To install the Ops Agent, open the instance template for editing, expand the Advanced options section, go to the Management field, and paste one of the following scripts:

      • For Linux OS:

        curl -sSO https://dl.google.com/cloudagents/add-google-cloud-ops-agent-repo.sh
sudo bash add-google-cloud-ops-agent-repo.sh --also-install

      • For Windows OS:

        Invoke-WebRequest -Uri https://dl.google.com/cloudagents/windows/install-google-cloud-ops-agent.ps1 -OutFile install-google-cloud-ops-agent.ps1
.\install-google-cloud-ops-agent.ps1

      For Ops Agent configuration details, see the Google Cloud documentation.

    • Legacy monitoring agent (not recommended)

      For the latest information about the legacy monitoring agent, see the Google Cloud documentation. Turbonomic aligns with Google Cloud's level of support for the agent.

  • Virtual CPU (vCPU)

    Virtual CPU is the measurement of CPU that is in use.

    For a managed or unmanaged instance group, vCPU represents the CPU used by all VMs in the group.

    Turbonomic calculates CPU based on the normalized CPU frequency and the number of vCPUs for a given VM. Normalized CPU frequency takes into account performance variations seen in different models of a given CPU platform. Because frequency is normalized, charts might show utilization values that are slightly higher than 100% (for example, 100.03%) when capacity is fully utilized.

  • Net throughput

    Net throughput is the rate of message delivery over a port.

    For a managed or unmanaged instance group, net throughput represents the average net throughput for all VMs in the group.

  • Storage access (IOPS)

    Storage access, also known as IOPS, is the per-second measurement of read and write access operations on a storage entity.

    For a managed or unmanaged instance group, IOPS represents the average IOPS for all VMs in the group.

  • I/O throughput

    I/O throughput is the measurement of an entity's throughput to the underlying storage.

    For a managed or unmanaged instance group, I/O throughput represents the average I/O throughput for all VMs in the group.

    For both storage access (IOPS) and I/O throughput, Turbonomic calculates capacity or uses capacity data published by Google Cloud, depending on the VM's machine type and disk.

    • Shared-core machine types share a physical core and are used for running small, non-resource intensive apps.

      For shared-core machine types with standard disks or SSDs, Turbonomic calculates capacity using internal benchmark data, which takes into consideration the observed maximum limit that can be achieved for IOPS and I/O throughput. It then uses the calculated capacity to analyze utilization more accurately.

      The use of internal benchmark data could result in differences in the capacity data shown in Turbonomic and Google Cloud.

    • For machine types that are not shared-core:

      • Turbonomic uses published capacity data and assumes that I/O block size is 16KB per I/O.

      • For machine types with persistent disks, Turbonomic assumes that the published capacity for the SSD disk type also applies to the balanced and extreme disk types. When a VM is attached to at least one of these disk types, capacity is assumed to be the per-VM limit for the SSD disk type. When a VM is attached only to the standard disk type, capacity is the per-VM limit for the standard disk type.